Experimental Design and Analysis for Tree Improvement provides a set of practical procedures to follow when planning, designing and analysing tree improvement trials.
Using many fully-worked examples, it outlines how to: design field, glasshouse and laboratory trials; efficiently collect data and construct electronic data files; pre-process data, screening for data quality and outliers; analyse data from single and across-site trials using either GenStat or SAS; and interpret the results from statistical analyses.
The authors address the many practical issues often faced in forest tree improvement trials and describe techniques that will give conclusive results with the minimum expense. The techniques provided are applicable to the improvement of not only trees, but to crops in general.
Building on the success of the first edition, this new edition has been fully revised and updated to relate to the latest commercially-available software packages for design generation (CycDesigN) and data pre-processing and automated generation of programs for statistical analysis (DataPlus). For analysis, it now provides both GenStat and SAS programs as generated by DataPlus.
Preface to the second edition
2 Experimental planning and layout
2.2 Experimental objectives
2.3 Sampling strategies
2.4 Allocation of resources
2.6 Analysis of variance
3 Data collection and pre-processing
3.2 Pre-processing using GenStat
3.3 Pre-processing using DataPlus
4 Experimental design
4.2 Simple designs
4.3 Factorial designs
4.4 Split-plot designs
5 Analysis across sites
5.2 Complete two-way tables
5.3 Incomplete two-way tables
5.4 Joint regression analysis with complete two-way tables
5.5 Joint regression analysis with incomplete two-way tables
6 Variance components and genetics concepts
6.2 Variance components
6.3 Genetics concepts
6.5 Heritability from provenance/progeny trials
7 Incomplete block designs
7.2 Need for incomplete block designs
7.3 Choice of incomplete block designs
7.4 Generalized lattice designs and alpha designs
7.5 Extra blocking structures
7.6 Using CycDesigN
8 Analysis of generalized lattice desings
8.2 Fixed-effects model
8.3 Mixed-effects model
8.4 Summary of analysis options
8.5 Nested seedlot structures
Appendix A Introduction to GenStat
A.2 GenStat commands
Appendix B SAS programs
“This is a well-produced book which achieves its aim of providing a set of practical procedures to follow when planning, designing and analysing tree improvement trials.”
Gillian M Arnold, University of Bristol (Journal of the Royal Statistical Society, Series A, v.157 no.2, May 2004)
Dr Emlyn Williams, a statistician with over 30 years research experience, is at the forefront of statistical research into the construction of efficient experimental designs, particularly for use in field, glasshouse and laboratory trials.
Dr Colin Matheson and Dr Chris Harwood and are internationally recognized tree breeders with extensive experience in developing, implementing and managing breeding programs for a wide range of Australian native and introduced tree species.
The authors, all from CSIRO Forestry and Forest Products, have over 20 years experience in several countries in the various activities involved in design and analysis of field trials. For many years they have been involved in training courses based around the book’s contents, held both in Australian and overseas